Development and characterization of an automated imaging workflow to generate clonally-derived cell lines for therapeutic proteins

SPEED-MODE cell line development (CLD): Reducing Chinese hamster ovary (CHO) CLD timelines via earlier suspension adaptation and maximizing time spent in the exponential growth phase

Chinese hamster ovary (CHO) cells are the preferred system for expression of therapeutic proteins and the majority of all biotherapeutics are being expressed by these cell lines. CHO expression systems are readily scalable, resistant to human adventitious agents, and have desirable post-translational modifications, such as glycosylation. Regardless, drug development as a whole is a very costly, complicated, and time-consuming process. Therefore, any improvements that result in reducing timelines are valuable and can provide patients with life-saving drugs earlier. Here we report an effective method (termed SPEED-MODE, herein) to speed up the Cell line Development (CLD) process in a targeted integration (TI) CHO CLD system. Our findings show that (1) earlier single cell cloning (SCC) of transfection pools, (2) speeding up initial titer screening turnaround time, (3) starting suspension adaptation of cultures sooner, and (4) maximizing the time CHO cultures spend in the exponential growth phase can reduce CLD timelines from ~4 to ~3 months. Interestingly, SPEED-MODE timelines closely match the theoretical minimum timeline for CHO CLD assuming that CHO cell division is the rate limiting factor. Clones obtained from SPEED-MODE CLD yielded comparable titer and product quality to those obtained via a standard CLD process. Hence, SPEED-MODE CLD is advantageous for manufacturing biotherapeutics in an industrial setting as it can significantly reduce CLD timelines without compromising titer or product quality.

Research on cell detection method for microfluidic single cell dispensing

Single cell dispensing techniques mainly include limiting dilution, fluorescent-activated cell sorting (FACS) and microfluidic approaches. Limiting dilution process is complicated by statistical analysis of clonally derived cell lines. Flow cytometry and conventional microfluidic chip methods utilize excitation fluorescence signals for detection, potentially causing a non-negligible effect on cell activity. In this paper, we implement a nearly non-destructive single-cell dispensing method based on object detection algorithm. To realize single cell detection, we have built automated image acquisition system and then employed PP-YOLO neural network model as detection framework. Through architecture comparison and parameter optimization, we select ResNet-18vd as backbone for feature extraction. We train and evaluate the flow cell detection model on train and test set consisting of 4076 and 453 annotated images respectively. Experiments show that the model inference an image of 320 × 320 pixels at least 0.9 ms with the precision of 98.6% on a NVidia A100 GPU, achieving a good balance of detection speed and accuracy.

High-throughput and automation advances for accelerating single-cell cloning, monoclonality and early phase clone screening steps in mammalian cell line development for biologics production

Mammalian cell line development is a multistep process wherein timelines for developing clonal cells to be used as manufacturing cell lines for biologics production can commonly extend to 9 months when no automation or modern molecular technologies are involved in the workflow. Steps in the cell line development workflow involving single-cell cloning, monoclonality assurance, productivity and stability screening are labor, time and resource intensive when performed manually. Introduction of automation and miniaturization in these steps has reduced the required manual labor, shortened timelines from months to weeks, and decreased the resources needed to develop manufacturing cell lines. This review summarizes the advances, benefits, comparisons and shortcomings of different automation platforms available in the market for rapid isolation of desired clonal cell lines for biologics production.

Microfluidic chip-based single-cell cloning to accelerate biologic production timelines

Cell line development (CLD) represents a critical, yet time‐consuming, step in the biomanufacturing process as significant resources are devoted to the scale‐up and screening of several hundreds to thousands of single‐cell clones. Typically, transfected pools are fully recovered from selection and characterized for growth, productivity, and product quality to identify the best pools suitable for single‐cell cloning (SCC) using limiting dilution or fluorescence‐activated cell sorting (FACS). Here we report the application of the Berkeley Lights Beacon Instrument (BLI) in an early SCC process to accelerate the CLD timeline. Transfected pools were single‐cell cloned when viabilities reached greater than 85% or during selection when viabilities were less than 30%. Clones isolated from these accelerated processes exhibited comparable growth, productivity, and product quality to those derived from a standard CLD process and fit into an existing manufacturing platform. With these approaches, up to a 30% reduction in the overall CLD timeline was achieved. Furthermore, early process‐derived clones demonstrated equivalent long‐term stability compared with standard process‐derived clones over 50 population doubling levels (PDLs). Taken together, the data supported early SCC on the BLI as an attractive approach to reducing the standard CLD timeline while still identifying clones with acceptable manufacturability.

Retrospective assessment of clonal origin of cell lines

Cell lines used for the manufacture of recombinant proteins are expected to arise from a single cell as a control strategy to limit variability and ensure consistent protein production. Health authorities require a minimum of two rounds of limiting dilution cloning or its equivalent to meet the requirement of single cell origin. However, many legacy cell lines may not have been generated with process meeting this criteria potentially impeding the path to commercialization. A general monoclonality assessment strategy was developed based on using the site of plasmid integration for a cell's identity. By comparing the identities of subclones from a master cell bank (MCB) to each other and that of the MCB, a probability of monoclonality was established. Two technologies were used for cell identity, Southern blot and a PCR assay based on plasmid-genome junction sequences identified by splinkerette PCR. Southern blot analysis revealed that subclones may have banding patterns that differ from each other and yet indicate monoclonal origin. Splinkerette PCR identifies cellular sequence flanking the point(s) of plasmid integration. The two assays together provide complimentary data for cell identity that enables proper monoclonality assessment and establishes that the three legacy cell lines investigated are all of clonal origin.

Developing therapeutic monoclonal antibodies at pandemic pace

The time from discovery to proof-of-concept trials could be reduced to 5–6 months from a traditional timeline of 10–12 months.

Single-cell dispensing and 'real-time' cell classification using convolutional neural networks for higher efficiency in single-cell cloning

Single-cell dispensing for automated cell isolation of individual cells has gained increased attention in the biopharmaceutical industry, mainly for production of clonal cell lines. Here, machine learning for classification of cell images is applied for ‘real-time’ cell viability sorting on a single-cell printer. We show that an extremely shallow convolutional neural network (CNN) for classification of low-complexity cell images outperforms more complex architectures. Datasets with hundreds of cell images from four different samples were used for training and validation of the CNNs. The clone recovery, i.e. the fraction of single-cells that grow to clonal colonies, is predicted to increase for all the samples investigated. Finally, a trained CNN was deployed on a c.sight single-cell printer for ‘real-time’ sorting of a CHO-K1 cells. On a sample with artificially damaged cells the clone recovery could be increased from 27% to 73%, thereby resulting in a significantly faster and more efficient cloning. Depending on the classification threshold, the frequency at which viable cells are dispensed could be increased by up to 65%. This technology for image-based cell sorting is highly versatile and can be expected to enable cell sorting by computer vision with respect to different criteria in the future.

Cyto-Mine: An Integrated, Picodroplet System for High-Throughput Single-Cell Analysis, Sorting, Dispensing, and Monoclonality Assurance

The primary goal of bioprocess cell line development is to obtain high product yields from robustly growing and well-defined clonal cell lines in timelines measured in weeks rather than months. Likewise, high-throughput screening of B cells and hybridomas is required for most cell line engineering workflows. A substantial bottleneck in these processes is detecting and isolating rare clonal cells with the required characteristics. Traditionally, this was achieved by the resource-intensive method of limiting dilution cloning, and more recently aided by semiautomated technologies such as cell sorting (e.g., fluorescence-activated cell sorting) and colony picking. In this paper we report on our novel Cyto-Mine Single Cell Analysis and Monoclonality Assurance System, which overcomes the limitations of current technologies by screening hundreds of thousands of individual cells for secreted target proteins, and then isolating and dispensing the highest producers into microtiter plate wells (MTP). The Cyto-Mine system performs this workflow using a fully integrated, microfluidic Cyto-Cartridge. Critically, all reagents and Cyto-Cartridges used are animal component-free (ACF) and sterile, thus allowing fast, robust, and safe isolation of desired cells.

High-throughput screening of antibody-expressing CHO clones using an automated shaken deep-well system

Biopharmaceutical protein manufacturing requires the highest producing cell lines to satisfy current multiple grams per liter requirements. Screening more clones increases the probability of identifying the high producers within the pool of available transfectant candidate cell lines. For the predominant industry mammalian host cell line, Chinese hamster ovary (CHO), traditional static‐batch culture screening does not correlate with the suspension fed‐batch culture used in manufacturing, and thus has little predictive utility. Small scale fed‐batch screens in suspension culture correlate better with bioreactor processes but a limited number of clones can be screened manually. Scaled‐down systems, such as shaken deep well plates, combined with automated liquid handling, offer a way for a limited number of scientists to screen many clones. A statistical analysis determined that 384 is the optimal number of clones to screen, with a 99% probability that six clones in the 95th percentile for productivity are included in the screen. To screen 384 clones efficiently by the predictive method of suspension fed‐batch, the authors developed a shaken deep‐well plate culturing platform, with an automated liquid handling system integrating cell counting and protein titering instruments. Critical factors allowing deep‐well suspension culture to correlate with shake flask culture were agitation speed and culture volume. Using our automated system, one scientist can screen five times more clones than by manual fed‐batch shake‐flask or shaken culture tube screens and can identify cell lines for some therapeutic protein projects with production levels greater than 6 g/L. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 34:1460–1471, 2018